Towards a model of constructional meaning for natural language understanding

Few researchers in natural language processing are nowadays concerned with linguistically-aware applications. On the contrary, the prevailing trend is towards the search of engineering solutions to practical problems, where researchers are motivated by the immediate gratification from the stochastic paradigm. As a result, there have been few attempts to confront the new challenges in linguistics from the natural language processing approach. The goal of this chapter is to introduce the theoretical foundation underlying ARTEMIS, a knowledge-based system which is intended to simulate natural language understanding in the framework of Role and Reference Grammar. More specifically, we will focus on how to enhance this functional model in order to make argumental constructions play a decisive role in the computational analysis of the deep semantics in the text.